[KidKhaos]

Good evening all!
I am planning out my next project and have decided I am building the rat rod of vertical machining centered.
First I was considering buying an X2 style mini mill. I then spotted a smoking deal for a RF-30 style machine and jumped on it without hesitation. After loading it up, on the way home I started putting thought into this machine and realized it is a terrible candidate for CNC being a column type machine you could only use the quill for a z axis. My suspicions were confirmed when I arrived home and realized people have the exact issue I feared. The machine is an excellent addition to my garage so I decided to keep it for light manual work and accurate drilling operations and move on to another platform. Again I was looking at the bench top options and had settle on a Precision Mathews when I realized I was paying $2500 for a dovetail machine then most likely another 2 grand on the low end to outfit her with CNC gear.
It was at this point I was feeling a little frustrated. I hate paying for an expensive thing and then replacing almost every part of it. It seems silly to tear apart a perfectly good machine and replace brand new parts. I could toss them on a shelf and wait for the day I never used them again.
I start to price steel for a more traditional type of build. Some quick math brought me to a total of 10k, maybe more. Wasn't I just looking at a $600 machine to outfit with $400 worth of electronics and another $200-$500 worth of random hardware into a itty bitty table top machining center?
Then it hits me. Man there are a lot of big chunks of structural steel out there no one wants that can be had for less than the price of scrap. Why not build a frame that is total overkill, outfit it with decent linear rails forgoing dovetail ways, and put some electronics on it I am familiar and comfortable operating?
So here is the plan. I Beams! They are everywhere and incredibly cheap. I have some pals that can probably straighten out the rail mounting surfaces once I get done warping the daylights out of them with my old Airco TIg/stick welder (Lets stick weld a precision machine!!!!). I will use mostly Chinese linear motion stuff. I have a sizable router table that was built from 80/20 profiles and Chinese SBR20 rails I can easily machine steel on with the right feed/speed recipe, so HGR style rails should do the trick. I have been looking at the different rails and see them up to and in excess of 45 mm width. BUT they get exponentially more expensive which I am assuming is due to the much lower demand for heavy duty rails. What if I were able to double up some smaller rails such as HGR15's as opposed to going with bigger rails? So in place of the traditional 2 rows of linear rails I would use four rows of smaller rails. I know it will require extra careful layout to get everything running straight, but it seems more than doable.
I guess my question is will four rows of 15mm linear rails be enough to machine steel on a small and rigid platform? Or should I buck up the cash and buy some manly 35-45mm rails and call it a day? For a fraction of the cost I might just be willing to try doubling down!

[joeavaerage]

Hi,
I made a small mill, 200mm x 200mm x 200mm machining volume, on cast iron beds with 15mm Rexroth rails, two per axis.
They are sufficiently rigid for my machine to mill steel.

My new project, a bigger version of the same thing, 350mm x 350mm x 350mm machining volume, cast iron beds, 32mm C5 ground
ballscrews with pairs of 20mm THK HSR rails and 750W Delta B2 AC servos. I anticipate the 20mm rails will be adequate.

In lots of respects the linear rails are easy, you can buy whatever size you need. Finding good ballscrews (C5 or better) is a much, much harder
battle, yet the ballscrews really set the tone of the machine rigidity and accuracy. Rather, finding good ballscrews is easy IF you pay big bucks for new
but trying to find GOOD second hand is much harder. I discounted ALL rolled C7 ballscrews as being inadequately accurate.

Craig

[maxspongebob]

Getting flat and parallel surfaces will be the hardest part of the build. Especially once you have warped it by welding. It is always amazing to me how much welding causes the metal to deform. Yes, you can use many techniques to minimize distortion, but it is never perfect.

I would recommend that you do a test. Determine how big the table will be and the spacing of the rails. Then go to the linear rail specs to determine how flat and how parallel your surfaces HAVE to be. Then weld up a small simple version of your VMC to see if you can actually make the surfaces meet the rail specs.

Also, more smaller rails seems like it would be more difficult to me.

[KidKhaos]

Getting flat and parallel surfaces will be the hardest part of the build. Especially once you have warped it by welding. It is always amazing to me how much welding causes the metal to deform. Yes, you can use many techniques to minimize distortion, but it is never perfect.

I would recommend that you do a test. Determine how big the table will be and the spacing of the rails. Then go to the linear rail specs to determine how flat and how parallel your surfaces HAVE to be. Then weld up a small simple version of your VMC to see if you can actually make the surfaces meet the rail specs.

Also, more smaller rails seems like it would be more difficult to me.

I agree! Fortunately I spent nearly ten years as a fabricator in a shop that rebuilt locomotive engines. The vast majority of my responsibilities included repairing major structural damage in engine blocks and oil pans with out warping the daylights out of them. So I have a lot of experience in these types of processes and feel confident I can weld up a decent frame without destroying it's integrity in the squareness department.
And I can probably coherse my old colleges to square up my new contraption on their massive mill once I have gotten her all welded up.

[maxspongebob]

Then I envy your access to those marvelous machining toys.

You didn't say what the envelope size is that you are looking to build. An RF-30 is rather small in the Y Axis. As for rail size, I think 35-45mm is huge. I have seen several good size project mills that used 20-25mm rails that worked out well.

I say go big or go home.

[peteeng]

Hi KK - 20mm rail will be fine for your machine. 25mm may give you some more room for your ballscrews. One thing about I beam is that it may vibrate, but make it big and I agree BE BIG or GO HOME> Peter

[joeavaerage]

Hi,
I looked at and seriously considered using heavy UC section steel (310 x 158). This would put the rail spacing about 300mm, the flange thickness
is 25mm so 20mm rail would be an easy fit. I require three beds, one for each axis. It was never my intent to weld these beds.
I had intended to make, probably by welding, Right angle frame to which the beds mounted.

In the event I am having beds cast in grey iron. There is just no beating cast iron when it comes to rigidity and stability. Naturally you can of course determine the dimensions
you want as opposed to those which are available by recycling structural steel. Also the vibration damping of cast iron is enviable.

The frame that mounts the beds is still in the design phase. A welded or otherwise fabricated steel frame is appealing. However I would prefer to cast a frame
in SG iron, being very tough, strong and stable. I suspect that the pattern and mold for such a large right angle frame all the webs etc will be expensive,
as much or more than the pouring of it. It will probably weigh in the region of 150kg.

The design will have to wait until the new year.

Today I took delivery of a 36 X 24 Grade B granite surface plate and installed it on the table I made for it. I already have a AA grade angle (only 6 x 6). I'm well on my way
to securing the means to make my machine flat and square.

Craig

[peteeng]

Hi Joe - A heavily bolted frame would be good. Adjustable, damp and stiff. Rivetted structures have stood the test of time vs welded structures and high strength bolts have overtaken rivets... Heavy UC with right angled brackets would be awesome...I'm looking at getting a section roller so I can radius the corners...Peter

[joeavaerage]

Hi,
absolutely no doubt that bolted steel structures are strong......but rigid is slightly different. The two are related but rigidity is a different
property.

How many times have you seen posts along the lines 'my new CNC machine flexes when cutting......it's as strong as hell...I don't understand'.
My first experiences were of that type.....but I learnt.

As you can see in my previous post I have got some great THK ballscrews, heavy duty new-old-stock 20mm THK rails/cars and what is not pictured
are three brand new Delta B2 series (17 bit incremental) 750W AC servos and drives, one with a brake, for the Z axis. I have not compromised any
component so far.....so why would I start compromising on the axis beds and/or the frame? Budget may preclude me from casting a frame as I wish
but I do have budget to cast the beds.

My previous mill, a smaller version of the same thing, I made the beds out of cast iron elevator weights. It worked out pretty good. I got each weight
(900x200x100 approx) for scrap value. The only problem is the quality of the cast was iffy, they had occasional inclusions and the surface 'rime' blunted
a few tools trying to peel it off. Should have expected it, I mean all a cast iron elevator weight has to be is heavy....quality is not an issue.
This time I'm getting my local foundry to do it....and I have seen their work, all very good, old school quality.


Craig

[KidKhaos]

I think you guys are right about two important aspects here.

1) Steel is springy stuff and transfers a vibrations as good as it does any other energy.

2) build volume: go big, or go home.

To address issue number one I have shifted my focus from an all steel frame to a hybrid steel and "epoxy granite" machine. While I like the idea of having something cast by a foundry, it isn't DIY enough for me. I really enjoy the build process. Maybe more than using the actual machine. Epoxy granite machines seem to have great vibration damping, AND rigidity. or maybe a decent compromise between the two. This method will also enable me to form some pretty complex and large components more easily. I have ordered a couple gallons of epoxy resin and have all sort of play sand around. I might upgrade some parts of my cnc router to experiment with this method. If I am being honest here though, I will probably just jump in feet first and start building a new machine. That is once I have addressed the second issue here.

Build volume. Shoot its easy to go straight gangbusters here and start buying 4000 mm linear rails... but how much is too much? I suppose my limiting factors would be machine foot print/height, and power ( stepper power and power supply to machine). First the physical volume isn't a huge deal as I am a one man household with a decent sized garage. My garage is somewhere in the 25 x 25 neighborhood, although I would prefer this thing to not take up a huge portion of said garage as I have a few other toys that like to call this same structure home. The second limiting factor is power. I am planning to use NEMA 34 closed loop steppers (orrrrrr maybe open loop with linear scales as encoders). as I increase the build volume I will also increase how much mass these bad boys will have to throw around which at some point will start to limit machine speed. I would prefer to over engineer in the stepper department and stay as far away from them losing steps as possible, while keeping what I would consider a reasonable machine speed. I learned on my first CNC router build there is nothing worse than a painfully slow machine. To be quite honest I have no idea how to calculate how much mass is too much, nor do I plan to. So in this aspect I will probably air to the side of caution and keep the build volume smaller than the top fueled maniac inside me screams for. I also have a CNC router that has about a 3' x 4' work volume and handles aluminum like a champ for the most part. Assuming I use due care and caution while calculating tool paths.
My other issue here is I started this out looking at buying a harbor freight bench top mill and converting it. Yet I sit here planing to build a friggen monster machine as I just can't help myself but live in excess one way or another. What do you guys thing? 20x16? I mean a PM-30 is around 30x18. I wouldn'y be mad about something like a 30x20 even. I could use multiple setups that way and machine cylinder heads if I wanted to. I am leaning toward the 30x20 here...

A couple other thoughts. I will be enclosing this machine as I would like to contain the chips and coolant. That will undoubtedly increase the overall size of this thing. Power supply to the machine isn't a huge deal as long as I stick with single phase. My garage is in desperate need of some more angry pixies anyway. My brother and I are getting ready to run a 200 amp supply to the garage while upgrading my house panel as well. My heater keeps tripping the main breaker in my house and that is super annoying.

Also, thanks for the feedback. I am greatly appreciative for it!

[joeavaerage]

Hi,
you want good....get AC servos...end of story.

While I like the idea of having something cast by a foundry, it isn't DIY enough for me

You make the patterns, they do the molds and pouring.

By the time you have invested in rails/ballscrews/servos you do not want to mess around with anything half-arsed.

Remember the song......'Accentuate the positive, Eliminate the negative, and don't mess with Mr. In Between'.
That means servos and cast iron.

Craig

[joeavaerage]

Hi,
as far as the size of the machine goes it will be determined by the ballscrews you can get.

If you want good then I assume you want ground C5 or better. Try to find those in your budget, they 'set the tone' of your machine,
size and accuracy. Everything else is measured in relation to those ballscrews, the size of the rails will complement the ballscrews,
the servos/steppers will be sized to get the resolution and speed you require.

Its likely that even a second hand ballscrew and support blocks will cost as much as a new Delta servo....so by the time you've paid
$400-$500 each for a ballscrew......what other components do you put with it?

Craig

[peteeng]

Hi KK - The ultimate design is driven by two factors its footprint and its working envelope. So decide on both then try to design a machine that fits this. Often the working envelope requires a bigger footprint then you think and vis versa. Once you compromise these in basic CAD world, the answer will settle out like gold in a pan. Then you can get serious with the other design points such as motors etc. You can't choose motors until you have settled the structures, so you know the weight of things. Its called a design spiral. You spiral down the options until it all gels in the end. Plus don't buy anything until you have resolved it in CAD then costed it. Then you have a good picture of where you are going... Peter

[KidKhaos]

A guy in my area is selling some Haas vf2 ballscrews for a really fair price. Maybe I'll check those out this weekend.

[joeavaerage]

Hi,
why is he selling them?

Have a look at these:
discounted automation products: NSK Precision Ground & Rolled Ballscrews, NSK Ball Screws: 25mm ⌀ thru 32mm ⌀

Double nut 10mm lead, 25mm diameter C5's with 890mm travel, new old stock for $395USD. That pretty good value, and 10 in stock.

Craig

[ihavenofish]

When you don't have a clue where to start, look at a machine you can buy that can do what you want.

For example, I have a brother TC229 tapping centre, with capabilities as follows:

450x300mm xy travel
1400ipm rapids
1.2 second tool change
able to drill 1" in steel, rigid tap 16mm in steel, swing a 2" face mill in steel, mill steel with a 16mm end mill (so says the blurb in the manual with recommended feeds and speeds).

So lets look at the mechanical construction.
- ball screws: 20mm diameter in the x and y, 25mm in the z. 10mm pitch, single nut, preloaded, c0 ground.
- linear guides: 25mm in the x and y with 2 rails and 4 blocks each, 30mm in the z with 2 rails and 4 blocks.
- servos: 720w 3600rpm in the x and y, 1000w 3600rpm in the z.
- spindle: BT30, 6000rpm, 2 nose bearings, no cooling, 3.7kw ac servo motor.

Not very big or heavy duty sounding is it? Yet with modern cam strategies it will out cut most other small machines under $100,000.

Most of the reason for that is how they constructed the frame. The frame is over 3000lbs of iron - the whole machine is 3700lbs. The rails for the Y are spaced 17" apart from each other, the ones on the X are 10" apart. The 1.5" thick solid slab table is bolted to another 1.5" sub table. The rails on the column are a foot apart, and the blocks on each rails are over a foot apart. The column is over a foot wide and deep. A brand new brother speedio - one of the highest performing small machines available - isn't much more beefy than my older model and has a virtually unchanged basic construction.

The common mistake I see people get caught on with DIY machines is thinking throwing 45mm surplus rails will somehow make up for a flimsy machine frame. It wont. Likewise, 32mm BNFN ball screws are designed for a 12000lb HMC, not your little hobby mill.

Welding a machine frame can be done, my maho hmc had a welded base frame from 10mm plate steel. Most high end cnc routers are all welded steel. As long as you can stress relieve and machine them afterward, there is no problem. If you cant, then, id suggest using aluminium instead with selective use of durabar grey iron where it makes sense.

[ihavenofish]

A guy in my area is selling some Haas vf2 ballscrews for a really fair price. Maybe I'll check those out this weekend.

Are you building a machine as big as a vf2? Cause, that's really not what you should be looking at.

THK BNK1510-C5-G0. or NSK compact FA 1510. Start there. Those are the default standard stock screws for all industries. The THK is a better screw, higher quality, stronger, but a few hundred dollars more (new). The NSK is smaller and can fit tighter spaces. Make you machine able to accept both with minimal fuss and account for them in your budget. Then when the design is done, scour for surplus, or compatible screws from other brands, then hit misumi for whatever ebay doesn't have. I got my surplus NSK 1510 screws for $115us each, brand new. Got a 1505 for my Z for $79, also new surplus. If your machine will be very heavy and powerful or have long axes, then you can up to 20mm screws.

[joeavaerage]

Hi,

Likewise, 32mm BNFN ball screws are designed for a 12000lb HMC, not your little hobby mill.

I disagree, you may have noted the rigidity specification I posted earlier for the 32mm BNFN screw, 1100N/um or about 110kg force per um deflection.
My mill is designed to be consistent with that level of force. The beds are cast iron 50mm thick weighing around 110kg each despite being only 700mm
long.

The frame is still being designed, I have been sketching and building wooden models. I am anticipating that the right angle flanges will be 250mm apart,
150 mm deep and 50mm thick. With the webs I am anticipating an as cast weight of about 250 kg. I will be casting the frame in SG iron.

The ballscrews are about 680mm long overall but have a travel of 358mm. The size of my mill is defined by those travels, the machining volume will
be 350mm x 350mm x 350mm. The linear rails are 650mm long so that places the centers of the cars 220mm apart (300mm overall). Transverse distance,
that is the distance between the rails is also 220mm on centerline (240mm overall).

All three axes will have a 750W AC servo and a brake on the Z axis.

I anticipate an all up weight of about 750kg.

In researching and planning this build I have looked extensively for second hand or otherwise surplus ballscrews and have yet to find any along the
lines that you have described at those prices. I would most certainly taken them up if I could have found any C5 ballscrews at those prices.

As I have said before the ballscrews determine the size of the machine. OP had best do some research to find out what is readily available and at what
price.

Craig

[ihavenofish]

Hi,



I disagree, you may have noted the rigidity specification I posted earlier for the 32mm BNFN screw, 1100N/um or about 110kg force per um deflection.
My mill is designed to be consistent with that level of force. The beds are cast iron 50mm thick weighing around 110kg each despite being only 700mm
long.

My 12000lb maho hmc had 3210 screws with a specified 10000N cutting thrust while tossing a 660kg work piece on a 1000lb 4th axis rotary table around at 0.3G. I still have the x screw... maybe I should sell it to a guy making a little diy hobby mill :P

there are 2 criteria that define the correct ball screw.
1: you need to know what rpm it has to spin at what amount of travel.
2: you need to know what the load it will see is.

You want to choose the smallest, shortest screw that meets those criteria. Larger screws have more inertia, which means you in turn need a much larger motor (your 750w servo in direct drive is probably not best suited to a 32mm screw). Larger screws also need larger support bearings, which cost exponentially more. They also need larger mounting hardware which takes up a lot of space and cost. Bigger isn't strictly better, you want "correct" for the application.

You have 6nm at the spindle motor. On a 10mm cutter that's 1200n force, distributed across all axes at once. One axis wont see more than say, 800n alone. Add in acceleration forces of 500-1000n. Under pretty hard conditions in the middle of a "too heavy" rough cut, you could call it 2000n for safety.

Speed vs travel wise, NSK compact fa 1005, 1210, 1510, and 2010 all will spin at 5000rpm at 350mm travel within spec.
Loadings are 2930n for the 10mm (cutting it too close), 3200n for the 12mm (better), 5460n for the 15mm (should survive even the hardest crash that doesn't also destroy everything else), and 8790n for the 20mm (getting into overkill). to add it in, the 25mm screws would take 12800n.

The nsk compact fa is a super light duty screw. The THK BNK will have higher loading on the same sizes - 9000n for the 1510 G0 (preloaded) model with 190n/um rigidity. 190n/um means about 10 microns, or 0.0004" deflection under your heaviest crash cut and effectively zero on a finishing cut.

In researching and planning this build I have looked extensively for second hand or otherwise surplus ballscrews and have yet to find any along the
lines that you have described at those prices. I would most certainly taken them up if I could have found any C5 ballscrews at those prices.

As I have said before the ballscrews determine the size of the machine. OP had best do some research to find out what is readily available and at what
price.

Craig

32mm screws, no... but, again... that's not the size that should be on a featherweight short travel machine like you've described.

[mactec54]

32mm screws, no... but, again... that's not the size that should be on a featherweight short travel machine like you've described.

All Bridgeport CNC Machines From the Boss series up, have a short travel Y axis of 12" ( 304.8mm ) and a X axis of 18" ( 457.2mm ) all have 32mm C3 Ballscrews with 5mm to 10mm pitch Screws 750W Ac servos work very well with them at 1:1 Ratio or direct mount, so no problem with anyone using 32mm ballscrews if that is what they have

So if they can fit them in there build all is good